{
  "mcpServers": {
    "nse-bhavcopy": {
      "url": "https://mcp.nseindia.in/bhavcopy/cm/mcp",
      "transport": "streamable-http"
    },
    "cm-market": {
      "url": "https://mcp.nseindia.in/cmmkt/mcp",
      "transport": "streamable-http"
    }
  }
}

import asyncio
import json
import polars as pl
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client

async def fetch_reliance_quote_clean():
    url = "https://mcp.nseindia.in/cmmkt/mcp"
    print(f"Connecting to NSE MCP server at {url}...")

    async with streamablehttp_client(url) as (read, write, session_id):
        async with ClientSession(read, write) as session:
            await session.initialize()

            target_tool = "cm_get_stock_quote"
            arguments = {"symbol": "RELIANCE"}

            print(f"Calling tool '{target_tool}' with arguments: {arguments}")
            result = await session.call_tool(target_tool, arguments=arguments)

            # Extract raw text content
            raw_text = None
            for content in result.content:
                if content.type == "text":
                    raw_text = content.text
                    break

            if not raw_text:
                print("Error: No text data returned from the MCP tool.")
                return

            try:
                parsed_json = json.loads(raw_text)

                # If the response wraps data in a dictionary, structure it for Polars
                if isinstance(parsed_json, dict):
                    # Extract fields if they are nested inside 'stock' or similar keys
                    data_list = [parsed_json]
                else:
                    data_list = parsed_json

                df = pl.DataFrame(data_list)

                # Check if 'stock' column exists as a struct and unpack it safely
                if "stock" in df.columns:
                    df = df.unnest("stock")

                print("\n--- Unpacked Polars DataFrame ---")
                print(df)

                # Export to CSV (now flattened successfully)
                output_filename = "reliance_live_quote_flat.csv"
                df.write_csv(output_filename)
                print(f"\nSuccessfully exported flattened data to {output_filename}")

                return df

            except json.JSONDecodeError:
                print("Failed to parse response text as JSON. Raw output:")
                print(raw_text)

# Run directly in your interactive cell
await fetch_reliance_quote_clean()